1. Field of the Invention
The present invention relates to an image discrimination apparatus and an image discrimination method for determining whether diaphragmatic radiation has been performed when an autoradiograph is obtained by irradiating a subject with radioactive rays using a radiographic device with a radiation diaphragm function, whether a designated area in the autoradiograph is an irradiated area, and whether a plain image (an area wherein a radioactive ray does not pass through a subject but is directly projected onto a sensor, etc.) is present in the autoradiograph. In particular, the present invention pertains to an image discrimination apparatus and an image discrimination method for identifying an autoradiograph by using the occurrence frequency at the edge portion, of a density value which is determined by a characteristic value obtained for the complete autoradiograph.
2. Related Background Art
Currently, in consonance, for example, with current digital techniques, autoradiographs are digitized, predetermined image processing is performed for the obtained digital images, and the resultant images are displayed on CRTs or are printed.
In radiography, “diaphragmatic radiation,” the irradiation of only a required area 803 of a radiographic area 800 of a subject 801, i.e., as is shown in FIG. 1, is generally employed for humanitarian reasons and in order to eliminate scattering in an area 802 that is not required, and to prevent color contrast deterioration. Further, to perform predetermined image processing for a thus obtained autoradiograph, a process parameter is generally selected from among a distribution of density values in the irradiated image portion, and the image processing is performed in consonance with the selected processing parameters.
However, when a portion to be irradiated is not specified, a portion that is not needed may also be irradiated. Then, information concerning the unneeded portion, i.e., unneeded image information, must be employed to determine the processing parameters, and appropriate image processing can not be performed.
Therefore, the irradiated image portion is extracted from the autoradiograph, and only image information for a required portion of the irradiated image is thereafter employed to determine the processing parameters.
To extract an irradiated image portion, with one example method an image density value is used to perform a differential function, and the obtained differential value is employed to identify the edge portions (irradiation edges) of the irradiated area. According to another method, a linear approximation expression is used to obtain an approximate value for edge portions of an area other than the irradiated area, and the irradiation edge is identified by using the difference between the approximate value that is obtained and the actual density value.
These methods are based on the assumption that an image was obtained by performing a diaphragmatic radiation process. Therefore, as the pre-processing performed for these methods, a determination is made as to whether the autoradiograph was obtained by using diaphragmatic radiation (the image was subjected to diaphragmatic radiation, as is shown in FIG. 1) or without using diaphragmatic radiation (the image was not subjected to diaphragmatic radiation, as is shown in FIG. 2).
To determine whether diaphragmatic radiation was employed, one method calls for the average density value or the middle density value in the center of the image to be compared with the average density value at the edge of the image, and when the average density value at the edge of the image is equal to or smaller than a predetermined value, it is ascertained that the image was obtained by using diaphragmatic radiation.
With the above described conventional image discrimination methods, however, when an arbitrary subject is radiographed without diaphragmatic radiation being used and the radiographic area includes the edge portion of the subject, the average density value at the edge of the obtained image varies in consonance with the size of the image portion, which includes the edge of the subject, and with the transmittance of radioactive rays. Therefore, even when the image is obtained without using diaphragmatic radiation, it may erroneously be determined that diaphragmatic radiation was used.
FIG. 3 shows an image obtained by autoradiography of the front of the lungs, using an X-ray. The reference a denotes an entire image, and the reference b denotes an irradiated area which is directly irradiated by the X-ray. A black portion included in the area b is a portion where the X-ray is directly incident onto a sensor. The reference A denotes an area to be used to discriminate whether the diaphragmatic radiation is employed or not.
With a method for discriminating whether or not the diaphragmatic radiation is employed, according to whether the portion A is irradiated by the X-ray or not, it can not be discriminated to which extent the diaphragm is opened, while such the method can discriminate whether or not the diaphragm is employed. Thus, since rough information as to the irradiated area is not obtained, a processing for extracting the irradiated area should be performed over the entire area. This may result in a problem that many time is required to perform such the process. In addition, even if the irradiated area is extracted, inspection of the extracted irradiated area can not be performed since the outline of the irradiated area is unknown.
Further, as is shown in FIG. 3, when the area on the sensor that is directly irradiated by X-rays is small, for lungs, the change in the density will be greater at the periphery of the lungs than at the irradiation edge, and the periphery of the lungs will be erroneously extracted as the irradiated area.
In addition, when the irradiated area includes the edge of the image and diaphragmatic radiation was not employed, the average density at the edge of the image fluctuates in consonance with the size of the irradiated area, which includes the edge of the image, and with the transmittance of radioactive rays. Therefore, a problem that has arisen is that it may be erroneously ascertained that diaphragmatic radiation has been employed, even though it has not. Another problem that has arisen is that when the intensity of radioactive rays is low there is no difference in the densities at the center of the image and at its edge, so that it will be erroneously ascertained that diaphragmatic radiation has been performed, even though it has not.
Furthermore, as image processing for an autoradiograph, there is a gradation conversion process for performing gradation conversion in accordance with the density distribution for an original image that is radiographed. For example, as data for a feature, the maximum density value is extracted from the original image, and the density value for the original image is so changed that the maximum density value is set to a predetermined value.
Generally, “diaphragmatic radiation” for radiating only a required portion of a subject is employed for radiography. This is done for humanitarian reasons, and in order to eliminate scattering in an unneeded area and to prevent color contrast deterioration.
In FIG. 4 is shown an X-ray image of lungs that were radiographed from the front and for which diaphragmatic radiation was not employed, and in FIG. 5 is an X-ray image of lungs that were radiographed from the front and for which diaphragmatic radiation was employed.
In FIGS. 4 and 5, an area 601 is an irradiation area (sensor portion), and a shaded portion 602 in FIG. 4 is a portion (plain image) obtained by directly irradiating the sensor with X-rays. An area 603 in FIG. 5 is a portion that is directly irradiated with X-rays when diaphragmatic radiation is employed. A plain image is not present in the portion 603, but tends to appear when diaphragmatic radiation is not employed for a subject.
In order to perform the gradation conversion process for the X-ray images (original images) shown in FIGS. 4 and 5, first, a check is performed to determine whether a plain image is present in the original image to be processed. When a plain image is present, it is removed from the original image. Then, for example, the maximum density value in the remaining portion is extracted as the maximum density value for the lung area, and the density value for the original image is so changed that the maximum density value for the film is approximately 2.0.
However, for the above described gradation conversion process, even though no plain image is present in an X-ray image obtained when diaphragmatic radiation is employed (see FIG. 5), conventionally it may erroneously be ascertained that the lung portion in the X-ray image is a plain image. Therefore, data for the feature of an image (e.g., the maximum density value in the lung portion), which are required for the gradation conversion process, can not be obtained.